Forecasting the Number of Road Accidents and Casualties using Random Forest Regression in the Context of Bangladesh
Last few decades, road accident is one of the rising problems in Bangladesh. Therefore, forecasting the number of accident and its casualties is necessary for its mitigation. In this perspective, machine learning can play an important role to predict the number of road accidents and its casualties in Bangladesh. To carry out this research, we collect the dataset of road accidents from the Bangladesh Bureau of Statistics website which is authenticated by the government of Bangladesh. Forecasting is done for the six metropolitan areas of Bangladesh i.e. DMP (Dhaka Metropolitan Area), CMP (Chittagong Metropolitan Area), KMP (Khulna Metropolitan Area), RMP (Rajshahi Metropolitan Area), BMP (Barisal Metropolitan Area), and SMP (Sylhet Metropolitan Area). Here, Random Forest Regression is applied to predict the number of road accidents and their casualties. Firstly, the regression model is trained on the training dataset and then the prediction is done on the test data. After forecasting, the predicted results are compared with the actual results which indicate that the random forest regression is good enough to forecast in this context.
Forecasting, Machine Learning, Random Forest Regression, Road Accident, Casualties